https://bugs.freedesktop.org/show_bug.cgi?id=50931

--- Comment #11 from Roman Eisele <[email protected]> ---
This problem is not limited to <br/>, there is some general parser problem with 
empty = standalone (X)HTML tags which match the pattern <[A-Za-z]+/>, because:

  <hr>   is imported correctly
  <hr /> is imported correctly
  <hr/>  fails: is just ignored!

I can not remember any other empty (standalone) (X)HTML tag besides <br/> and
<hr/> which can occur without additional attributes. In empty = standalone
(X)HTML tags which *require* some attributes, like <img/>, there is no problem
with missing whitespace before the final />:

  <img src="x.png">   is imported correctly
  <img src="x.png" /> is imported correctly
  <img src="x.png"/>  is imported correctly

And as soon as I add some attribute to the <br/> and <hr/> tags, the problem
vanishes:

  <hr class="myhr">   is imported correctly
  <hr class="myhr" /> is imported correctly
  <hr class="myhr"/>  is imported correctly

So the problem is really limited to tags which match the pattern <[A-Za-z]+/>.

Therefore:

(In reply to comment #10)
> Is /core/svtools/source/svhtml/parhtml.cxx:1118 here to blame?

Seems reasonsable, but my code reading facilities are limited. Would it help if
one changes line 1118-1119 from

    } while( '>' != nNextCh && !HTML_ISSPACE( nNextCh ) &&
             IsParserWorking() && !rInput.IsEof() );

to something like

    } while( '>' != nNextCh && '/' != nNextCh && !HTML_ISSPACE( nNextCh )
             && IsParserWorking() && !rInput.IsEof() );

?

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to